This application is a renewal of the application titled ACTIVE Phase II: PSU Field Site. This application is for the Field Site at The Pennsylvania State University. Phase I of ACTIVE (Advanced Cognitive Training for Independent and Vital Elderly) was a randomized controlled trial of three cognitive intervention arms, addressing the question of whether improving basic cognition aided in maintaining functional independence in elders. Phase One found strong, broad and durable cognitive ability-specific training effects but minimal transfer to everyday activities (i.e., functional competence). Through the two-year followup, there was no evidence of a significant decline in ADL and IADL status. To adequately understand the cognitive transfer effects of the training interventions requires a longer followup period, particularly to see whether there is a separation of the change trajectories for everyday activities of trained and untrained participants over time. Phase II of ACTIVE is proposed as a followup study focused on measuring the longterm impact of training effects on cognitive function and cognitively demanding everyday activities. The Phase II followup will consist of one assessment to include the Phase I post-test battery and a clinical assessment. The ACTIVE cohort (n = 2832) is a special sample, containing substantial oversampling of African American, socioeconomically poor, and very old adults. The Specific Aims of Phase II of ACTIVE are: 1) to determine whether the cognitive interventions (as initial treatment or as a consequence of repeated boosters) have long-term protective effects on functional outcomes; 2) to document any delayed transfer of the cognitive training to secondary outcomes; and 3) to identify individual factors that affect response to intervention. As in Phase I, the primary analytical approach to detecting treatment effects on both cognitive and functional abilities will be a repeated-measures, mixed-effects model incorporating all design features as fixed effects and individual-level variability as random effects. Other multivariate analyses including lagged and cross-lagged analyses of change using latent change analysis, structuralmodeling, and growth curve analyses will also be used as appropriate to characterize relationships between individual difference factors and change in functional competence. Retention is projected conservatively at 72% with 65% of the cohort providing full data and another 7% providing partial data at year 5. Power analysis shows that extending the study will make it possible to observe effect sizes on the order of 0.05-0.10 with excellent power, in the range of at least 80-90%.